Hopstack, headquartered in Palo Alto, CA, is the provider of an advanced AI-powered warehouse management and execution SaaS platform to automate and digitize warehouse operations

Hopstack

Hello! We are developers from the Navigine company. We decided to test the Wi-Fi RTT technology and share the results with you. What is Wi-Fi RTT Wi-Fi RTT (Round-trip time) is a brand new technology that allows locating devices with unprecedented sub-meter accuracy. Basically, it measures the distance between a device and Wi-Fi routers which is way more accurate than using a GPS signal. The latter can’t provide an acceptable level of accuracy at all when it comes to indoor positioning. Wi-Fi RTT outperforms RSSI-based methods as well, thanks to reduced latency and better accuracy. The Use Today, this technology is the next step to improve services that use indoor positioning. We expect that its implementation will become a better solution for many industries. Wi-Fi RTT is applicable for tracking vehicles and cargo within warehouses and production plants. Comparing to BLE, this technology provides enhanced accuracy and reduced latency, making processes more rapid and reducing delays. It is effective for public hotspots as well, such as offices and malls, airports, exhibitions and not only. Indoor navigation is the solution users need for making their visits to public places more comfortable and enjoyable. Through this, they can plan their time effectively and find locations they need immediately. The Advantages Our solution goes further with unique algorithms that enhance the basic capabilities of Wi-Fi RTT. demonstrates a submeter accuracy in checkpoints 95% of the time and around 30 cm accuracy 50% of the time. The positioning engine is a particle filter combining Pedestrian Dead Reckoning (PDR), based on inertial sensors readings, and Wi-Fi RTT readings for correction. Meanwhile, the dynamic test shows that delay in position estimation does not exceed one second. The latter result is achieved by complementing Wi-Fi RTT measurements with PDR via particle filter. Semi-static test in our office The technology is not just an innovation in itself, but also a practical solution that improves business efficiency and customer experience. Wi-Fi RTT provides longer distance measurements and can be integrated into existing Wi-Fi network infrastructure. Current status Now, it is too early to refer to any cases, while we are currently working on raising experts interest from LinkedIn. We can see the real need for this technology, so the first implementations are coming soon in Q1 2020. Today, we have confirmed orders from a big carrier and one logistics company. Why Wi-Fi RTT? The industries using indoor positioning have to realize that often BLE is not the solution that meets their needs. Wi-Fi RTT is a significantly more relevant technology for many cases. Our solution based on Wi-Fi RTT delivers clear advantages for customers, including high accuracy, low latency, ease of use and integration. With just 2 weeks between order and integration, it is one of the most rapid and flexible solutions. The combination of high accuracy, availability, and low price, can make the technology an Indoor positioning market leader.

A Needle in a Haystack: How Wi-Fi RTT Takes Indoor Positioning to the Next Level

Finding enterprise use cases worth development dollars Earlier this year, the GSMA, the world’s most influential telecom consortium, published a paper detailing Some of them were compelling. Others, less so. fifteen blockchain use cases for mobile network operators . As an , I was happy to see this kind of attention given to my still-nascent field. (Is all press good press?) But as a founder whose ability to buy Chipotle burritos rests on how seriously and successfully telecom carriers adopt blockchain technology, the addition of those less compelling use cases had me worried. entrepreneur in the telecom and blockchain space The sad truth is that the blockchain world is rife with misinformation, poorly-thought-out applications, and centralized tech pretending it’s “de-”. If left to their own devices, I could imagine carriers attempting the use cases I saw as less promising and being forever turned off of blockchain. So, although it’s against my very nature as a native Midwesterner, I see no other alternative to raining on this particular parade. As Warren Buffett wisely said in his : '17 shareholder letter The less the prudence with which others conduct their affairs, the greater the prudence with which we must conduct our own. In this article, I’ll attempt to be a voice of prudence. The applications that the GSMA proposes range from great and world-changing to fundamentally flawed, and I’ll explore four of them in the sections that follow. In so doing, I’ll draw out lessons worth remembering not only in telecom but also in broader enterprise blockchain R&D. A quick note on methodology: as an IDEO CoLab Fellow last winter, I learned that new tech is most interesting when it is desirable (i.e., adds real value for real people), feasible (i.e., technologically possible), and viable (i.e., good business). Check out IDEO’s explanation here if you’re curious, but it’s easiest to just witness them in action. You’ll get it. #1: Smart contracts for 5G network management “To realise the 5G promise of ubiquitous access across various networks, CSPs will need to handle heterogeneous access nodes and diverse access mechanisms. Selecting the fastest access node for every user or machine will be a central challenge in the future. …Rules and agreements between the various access providing networks can be coded into autonomous smart contracts. These contracts can be dynamic in nature wherein any time a policy needs to be changed, only the contract code needs to be changed.” The GSMA’s first use case falls prey to a classic problem of outside-in entrepreneurship — it describes a solution that every consumer wants but no business will offer. Is it to connect to any network freely and get the best value for the best service? Of course! I don’t care who gives me four bars of LTE service as long as I can reliably stream the #7-ranked Michigan basketball game in the back of my Uber. desirable But from the carrier’s point of view, this solution is neither feasible nor viable. On the technological side, the use case makes a number of critical leaps into still-unknown areas of blockchain tech, but far more damning is the business critique: why would carriers ever give up control of their customers? Carriers have built empires by leveraging their network assets to attract customers and reduce churn. Would they really buy into a protocol that forces them to auction for every data session, text, and call? feasibility viability Sounds tautological, but it’s surprisingly easy to forget. Yes, real-time bidding for connectivity would be amazing. But if it ever exists, it will only be after a prolonged battle with operators the world over. The lesson here is for prospective entrepreneurs: keep the enterprise in mind when designing enterprise blockchain products. #2: IOT connectivity “While localised networks in for instance a manufacturing line can provide great benefits, self-organised global supply chains powered by IOT can help streamline processes that require a lot of paperwork and governance. This would shift the focus from centralised IOT platforms (in technology and governance) to self-governed mesh networks… IOT can also be used for automatic tracking for quality and conditional parameters during transports, switching of ownership and provenance use cases.” The GSMA offers up three use cases in the IOT space. Unfortunately, they all fall prey to the same problem — they’re smart, and should probably happen, but they probably shouldn’t happen on a blockchain. These use cases are : IOT management via blockchain is a use case well-explored by startups and large enterprises alike. They may even prove to be : IOT is a huge and growing industry, devices will get smarter and smaller over time, and managing devices efficiently is a huge, valid concern. feasible viable The problem arises, though, when we consider the of blockchain in this context given the problems we are asking it to solve. desirability As I understand it, the problem is that we can’t get good data about products flowing through a manufacturing line or a multinational supply chain, so we’re asking blockchain to help. Walmart, for instance, now mandates that farms that supply their stores leafy green vegetables enter “ ” that they’ve build in conjunction with IBM. detailed information about their food into a blockchain database But wouldn’t that “detailed information” be just as useful in a normal database? In my mind, this type of supply chain management is a poor use of blockchain. If you can’t trust your supply chain partners (or much less, the folks on your manufacturing line) to produce accurate data, it doesn’t matter whether you use a blockchain, a database, or a stone tablet to store it: your customers are still going to . Garbage in, garbage out. (Or, due to blockchain’s immutability, it’s more like garbage in, garbage forever.) get E. coli from your lettuce Blockchain is expensive, slow, and non-private. It costs significant time and resources to write to a blockchain. Barriers to entry are high. (Do Walmart’s farmers now have to own wallets or accounts with IBM?) The lesson here is never to use blockchain when a normal database will do. This is not to say that blockchain won’t have a place in IOT; I think it will. But I’m willing to bet that device security turns out to be a better use case than provisioning or self-maintaining supply chains, because the former is a better fit to the unique constraints and benefits of blockchain. #3: Fraud prevention The telecom industry has not yet found a way to effectively and sustainably prevent fraud. In a general sense, working together to combat interoperator fraud is a good way to start. Blockchain could for instance be used to govern access to fraud detection information shared between operators…another angle is fraud caused by delays in sharing [call detail records] between operators. As blockchain works in real-time, using a shared administration of network activity could help identity plausible fraudulent connections faster. Smart contracts could be deployed to actively enforce network usage compliance rules and help us resolve inter operator disputes easier. Surprise twist: I actually this upcoming use case! In fact, I like it so much that I’ve been working on it for the past year with . (Is that cheating?) like drect.ly From a standpoint, blockchain has some clear advantages over other solutions in fraud prevention. In fact, it’s so good that you get to (justifiably) recite the holy text of blockchain use cases: desirability This is a case when mutually-distrustful actors need to coordinate to maintain a shared version of the truth for an economically-valuable digital record. That sentence is why blockchain was created; in this use case, we’ve just replaced the economically-valuable digital record of bitcoin ownership to a record of which telecoms have provided how much coverage to which roaming customers. Those records, it turns out, are incredibly valuable (and therefore, if managed well, can sustain a business model). Every year, telecom carriers lose over $38 billion to fraud; even a relatively small dent in that number is large in absolute terms. viable My own incredible bias notwithstanding, I do recognize some challenges here with . Billions of records are exchanged daily, which poses scalability problems. These records expose significant, closely-held data about network usage and subscriber behavior, which poses privacy concerns. Governance is challenging, for it’s unclear how centralized my ownership of the code base makes this project. And how can you get bad actors to sign onto a platform that demands higher responsibility? feasibility All that said, these are solvable problems. And perhaps that’s another lesson: ensure that your use case is a great fundamental fit with the unique value prop of blockchain technology — and then prove the feasibility haters wrong. If you think about it, that’s really the only way that technology advances. any #4: Hosted services Distributed Ledger and/or Blockchain network operators are responsible for configuring the network and granting access to network participants. In general, they specify the use case(s) that the network is serving, and are often involved in managing the network in terms of software upgrades, arbitration services, etc. A network can be operated by a single entity or a federation of multiple, separate entities. As a strategy consultant at Deloitte, I helped write a . It generated over 100 million views before being lost to the sands of business-thought-pieces past, but I’ll always remember it as the first time that I learned John Hagel’s concept of innovation “ .” trends report on ecosystems thinking antibodies The SparkNotes version of Hagel’s idea is that antibodies (i.e., agents of resistance) arise in big organizations whenever change is coming. They fight for the preservation of the status quo that helped them rise to power, and, even if well-intentioned, they often stifle the speed of change by insisting that old rules, norms, and values govern fundamentally new contexts. For example, a new business unit might be shut down because it’s cannibalizing sales of the core product; that is judged by the old rule of “protect the core,” rather than the new reality that any player could release the cannibalizing product and it’s better to optimize for growth. In the blockchain context, antibodies take the form of centralized services masquerading as blockchain technology. In the old world, businesses want efficiency and control — but Proof of Work is deliberately inefficient, and blockchain networks are intentionally decentralized. When I read a use case like the “hosted services” from above, where blockchains are treated like permissioned databases, I can’t help but remember Hagel’s antibodies. The whole point of blockchain technology is to make public commitments that are impossible to reverse; to open up previously siloed networks; and to radically decentralize power, code, and economic value accrual. So, my final lesson is this: You’ll want to do so — believe me, as someone who has built these systems — but if you do, you’ll be further from the real promise of blockchain technology and closer to designing a really bad database. After living in a centralized world, decentralization feels innately wrong. But as we enter this new age of decentralized technology, we ought to remember that it’s okay to break all of the old rules. never surrender decentralization when designing blockchain use cases. In fact, that’s the whole point. To connect with Christian, reach out on Twitter ( @cdkeil ) or via email at christian [at] drect.ly.

Needles in the Blockchain Haystack

How do you find the job candidate who’s an exact fit for your organization? How do find the person you are going to marry? How do you locate the right house for you and your family? How about a pet that’s been missing? Worse yet, a child who’s been missing? As a society, we spend a lot of time looking for things. Individuals search for apartments to rent, cars to buy, and romantic partners to date. Businesses search for customers, employees, and suppliers. . Our goal is to enable people to use incentive programs and in order to solve some important problems facing the world today. We believe, in particular, that we can help people Although we tend to approach these searches as if they are unique, we at nCent believe that they appear distinct on the surface but are often subject to similar dynamics on a meta level blockchain technology find higher quality things in a fraction of the time and more efficiently than they might expect. Some search problems, like finding the closest Italian restaurant, are easy to solve. , for example, is more challenging because it may be unclear where to start and how to efficiently gather information. It’s obviously a high stakes situation, because it involves the welfare of a child — quite a bit different then satisfying a pasta hankering. But other search problems are more difficult and involve higher stakes. Finding a missing child Difficult searches are called problems. “needle in haystack” One of our guiding principles at nCent is our belief that these “haystack” problems can be optimally solved by combining decentralized networks, blockchain technology, and the right incentives programs. A ‘proof of concept’ for this approach happened in when DARPA (USA defense research agency) placed ten red balloons in secret locations around the U.S. and offered $40,000 to the first team to locate them all. The winning team solved this search problem by offering recursive incentives: Anyone who either correctly identified the location of a red balloon or someone whose information led to a person who identified the location of a red balloon was rewarded. For example, if Person A referred Person B who referred Person C who correctly identified the location of a red balloon, then all three people in this chain would get rewarded in some capacity. 2009 referred Even if you don’t know where the needle is in the haystack, you are still incentivized to help by referring others to get closer. If the tail end of your subtree finds the needle, you too will be rewarded. Everyone who contributes to the network can be absolutely sure that if their contribution is valuable, their reward will be commensurate with this value. This means that the person who finds the missing child the person who spread the word to the finder’s social media network will be rewarded. The person who refers a successful candidate for a specialized job the person who informed the referrer will be rewarded. These programs can reduce their overall cost with (Larger payouts for those who are “closer” in the chain to the final goal) Utilizing blockchain technology in this process enables the network to be transparent and to be audited by anyone at any time. and and recursive payout structures . These recursive programs are scalable and when designed correctly, have an upper bound on the cost of sponsoring them. Recursive incentives motivate people to inform other people of the search problem. When executed properly, these networks grow quickly and organically, increasing the statistical probability of finding the target. In fact, we believe that recursive incentives, where every node of the successful chain is rewarded, become increasingly more valuable as the search problem becomes more difficult. Spam is another common problem that we will address in a forthcoming post: The biggest hurdle to solving a haystack problem is growing the search network quickly enough so that the chance of identifying the needle increases exponentially. “Cashout: What Does it Mean to be Spammed with Money?” As the tail end of the network is further removed from the origin of the network, the spread of information will be more efficient than any other search process implemented in the marketplace. . The incentive structures implemented on our blockchain will be modular enough to apply to many types of search problems. The inclusion of “decentralized networks” and “blockchains” might sound like tech jargon, but in reality it’s just the underlying tech that enables you to give “asking your friends for help” a global reach! We fully recognize that each case will come with its own idiosyncrasies, ranging from economic constraints of the searcher to the psychological realities of how people respond to incentives To learn more about our work, email me directly at kk@ncnt.io . To stay in the nCent loop, hear me tweet and join our international telegram channel .

How We Can Find the Needle in a Haystack, Together

Pinpointing Root Causes of Application Performance Problems in Dynamic Environments Pinpointing the root cause of application performance problems in a dynamic environment often resembles a complicated game of hide-and-seek. A familiar analogy is that it’s like trying to find a needle in a haystack. This characterization illustrates the challenge but isn’t especially helpful in offering solutions. So, if the root cause of an application problem is the “needle” and the application environment is the “haystack,” how exactly would we go about restoring peak performance? Taking the analogy literally is actually a pretty good way to explore the various approaches to resolving application performance problems in dynamic, elastic, and transient infrastructure environments. So, exactly how would you go about finding a needle in a haystack? 1. Eliminate the haystack Probably the fastest way to find the needle is to burn the haystack. Or dumping the hay into a pool of water and waiting for the needle to sink to the bottom would also work. If we were really talking about needles and haystacks, fine — but we’re not. There are fundamental business reasons why more and more companies are turning to dynamic cloud-based deployments in the first place. We can pretty much reject the option of returning to monolithic, static infrastructure where identifying application performance cause-and-effect relationships is more straightforward. 2. Rummage for the needle manually We could sift through the pile of hay, strand by strand, until we find the needle. That’s basically the way traditional (APM) tools work. They throw a massive collection of data at us (albeit usually in nicely formatted charts and graphs) and require us to comb through it all to figure what went wrong. Actually finding the needle remains a laborious manual process. Somebody somewhere has to pore over the charts and graphs, make sense of the data, and connect the dots to discover the root cause. Such handwork analysis can take hours — even days. application performance monitoring That’s too long when application problems are leading to disgruntled customers. Of course, things would go faster if we assigned more staff to tackle the problem: the more people sifting through the hay, the sooner we’d find the needle. But convening a war-room assembly to interpret the data is still a labor-intensive, time-consuming effort. Already most CIOs are frustrated that so many man-hours are spent on troubleshooting instead of innovating and optimizing. Furthermore, and perhaps more significant, today’s liquid environments — where millions of dependencies are changing on the fly from moment to moment — are simply too complex and too fast-moving, with too many performance data points, for humans to analyze in a timely manner. It would take more people than any business has at its disposal to detect causal relationships among so much data and identify the component that’s causing the problem. 3. Locate the general area of the needle Using a metal detector or an X-ray would narrow down the approximate location of the needle within a haystack. This is more efficient than looking through the entire pile, but it’s still a lot of work. We still have to manually sift through the hay — just not as much. Similarly, a handful of modern APM solutions get us a step closer. Discovery agents collect and correlate myriad metrics and performance data points from various distributed elements across the application environment. The data is consolidated and presented in a unified, integrated “single pane of glass” view. We now have a clear, comprehensive picture of the of the performance problem, but we haven’t diagnosed the The hard work is still left to us. We still have to pore through charts and graphs of data — just not as much. Modern APM displays the data within a more meaningful framework that helps us narrow down our focus. symptoms root cause. Now, imagine the difficulty when the needle is constantly on the move. That’s exactly what’s happening with today’s highly dynamic environments. Servers are perpetually coming and going, auto-scaling up and down as needs dictate. Container technology and microservices architecture are even more transient — moving from host to host — and generate even more metrics. With our applications now having so many moving parts, trying to keep up with them and their topology by dissecting a plethora of charts and graphs is literally beyond human capabilities. Dynamic applications have millions of dependencies that need to be analyzed, and they’re ever changing. No sooner do we think we have a handle on how everything is working together than the landscape changes and we’re back to square one. Present-day application environments change too fast and too often for even the most talented troubleshooters. Even a relatively simple containerized environment like this one, with only 142 hosts, has more dependencies than a human being can track without software assistance. 4. Extract the needle automatically Ultimately, the best way to find the needle (especially if it won’t sit still) is to use a powerful magnet to pull it right out of the haystack and deliver it into your hands. In our analogy, a next-generation solution would be such a “magnet.” It zeroes in on an application problem’s root cause automatically, rendering time-consuming manual analysis unnecessary — the same way a magnet obviates the need to sift through the pile of hay. artificial intelligence–powered APM Artificial intelligence (AI) enables computers to do what they do best: absorb huge amounts of information and make sense of it faster and more thoroughly than is humanly possible. It applies sophisticated algorithms and context-rich diagnostics to detect and correlate causal relationships among data with millions of dependencies in microseconds. In essence, AI-enhanced monitoring analyzes the problem the same way we humans would, only faster. Artificial intelligence is able to perform almost instantaneously the analytical “heavy lifting” that takes us humans hours or days. From theory to reality: Dynatrace APM Artificial intelligence has been defined as “being able to perceive its environment and apply cognitive functions such as learning and problem-solving.” has been built on artificial intelligence so that it is able to Dynatrace APM auto-discover all components of the full technology stack end to end from the customers’ web browsers all the way down to the host infrastructure — map out the entire IT in an interactive visual display environment identify the millions of dependencies among websites, applications, services, processes, hosts, , and cloud infrastructure networks learn how it all works together and what constitutes normal behavior automatically detect, analyze, and prioritize anomalies and performance problems actually recommend solutions to the root cause of those problems And AI allows Dynatrace to do all this at a speed, scale, and precision that no human could ever match. Dynatrace gets to “know” applications. Its proprietary Smartscape technology continuously auto-discovers and monitors every aspect of every application. Sophisticated algorithms learn normal application performance patterns and proactively flag anomalies. The AI auto-adjusts baselines dynamically in real time to avoid false positives caused by preset static alerting thresholds. Application performance problems are seldom isolated, one-time events, and they’re usually symptoms of a larger issue. Dynatrace looks at all other transactions that used the same components around the same time to see if they also experienced problems. Artificial intelligence correlates events throughout the full technology stack — client side, server side, infrastructure level — to identify analogous issues and detect causal relationships. AI diagnostics analyze the vast amount of data collected to pinpoint the exact component that’s causing the problem. In today’s increasingly complex and fluid IT environments, interpreting data points with millions of dependencies in order to determine underlying casual relationships is more than humans can master by themselves. Only artificial intelligence can take the colossal volume of data and translate it into actionable solutions before problems hit customers. See for yourself what Dynatrace can do. Dynatrace offers a — the complete full-blown version, not a stripped-down demo — that auto-discovers and visually maps out your entire technology stack within minutes. free trial

Finding the Needle in a Haystack

Using the Stratification Method for the Experiment Analysis

Recently, Apple released the Vision Pro, and I was fortunate enough to become a proud owner. Since then, I've been deeply engaged in developing my application and exploring various apps and trends within the Vision Pro App Store. One thing that grabbed my attention is that new apps are coming out every day, but only a few fit well with the immersive virtual reality experience. I guess many developers struggle to find documentation and examples for Vision Pro features. However, utilizing the capabilities of Vision Pro can set your application apart and provide users with an enhanced experience. In this article, I'll be sharing some tips and tricks I've uncovered while crafting my application. These tips will not only help your application to stand out but also ensure its seamless integration and native feel within VisionOS. Dark Surroundings Effect I'm sure you've noticed that when viewing a video or photo, the background behind the window is darkened, in the settings it's called "auto-dimming". To get the same effect you need to use .preferredSurroundingsEffect(.systemDark) on your view and then the space around the user will darken and your window will remain bright, this will help to keep the focus on the content. Hidden Window Controls Line When viewing a video or content, you want the user to have no distractions and create the effect of a window floating in the air, for this Apple has given us the ability to hide the bottom bar of the window that is used to move or close the current window. With .persistentSystemOverlays(.hidden) you can hide this control bar on your window, it will disappear with the default animation and only appear when the user interacts with the window. Video Reflection Effect If you open the standard Apple TV application and turn on a movie, you will notice that the movie window has a reflection that changes with each frame, giving a more immersive effect. To get a glow effect around your video, you need to use VideoPlayerComponent and enable isPassthroughTintingEnabled, then the effect will be the same as in Apple TV or HBO Max. Here is what the documentation says: This is to enable passthrough tinting during video playback that shows up around the video taking the average color of the frame and tinting the passthrough with that color to emphasize the video. Window Resizing You may want to prevent the user from resizing the window, there is a way to do that: onAppear {
    guard let windowScene = UIApplication.shared.connectedScenes.first as?UIWindowScene else { return }
    windowScene.requestGeometryUpdate(.Vision(resizingRestrictions: UIWindowScene.ResizingRestrictions.none))
} You can also specify additional size settings in this method The full signature of the method looks like this: windowScene.requestGeometryUpdate(.Vision(size:, minimumSize:, maximumSize:, resizingRestrictions:)) Hands Visibility If your application uses immersive space and you don’t want the user to see their hands, or you want to replace them with virtual hands, as is done in AmazeVR, then you need to use .upperLimbVisibility(.hidden) Ornament I'm sure you've seen sidebars in VisionOS applications that look like TabBars. Apple has introduced a new View method for VisionOS that allows you to "expand" the window and add ornaments on either side of it. .ornament(
    visibility: .visible,
    attachmentAnchor: .scene(.bottom),
    contentAlignment: .center
) {
    HStack {
        Button("Play", systemImage: "play.fill") {
        }
        Button("Stop", systemImage: "stop.fill") {
        }
        
    }
    .labelStyle(.iconOnly)
    .padding(.all)
    .glassBackgroundEffect()
} Сore location An interesting fact that I was able to verify experimentally is that CMMotionManager's methods only work in immersive space and not in the normal window view. Conclusion VisionOS provides many tools to create immersive applications that bring users into the virtual world and make them feel part of it. While more and more applications are being released for Vision Pro, only a few utilize its full range of features. That's where your chance to distinguish comes in — use them and take advantage. Unfortunately, developers struggle to find easily accessible documentation and examples. That's why I'm writing this article. I hope my tips will help you make your application better. Recently, Apple released the Vision Pro, and I was fortunate enough to become a proud owner. Since then, I've been deeply engaged in developing my application and exploring various apps and trends within the Vision Pro App Store. Apple One thing that grabbed my attention is that new apps are coming out every day, but only a few fit well with the immersive virtual reality experience. I guess many developers struggle to find documentation and examples for Vision Pro features. However, utilizing the capabilities of Vision Pro can set your application apart and provide users with an enhanced experience. In this article, I'll be sharing some tips and tricks I've uncovered while crafting my application. These tips will not only help your application to stand out but also ensure its seamless integration and native feel within VisionOS. Dark Surroundings Effect I'm sure you've noticed that when viewing a video or photo, the background behind the window is darkened, in the settings it's called "auto-dimming". To get the same effect you need to use .preferredSurroundingsEffect(.systemDark) on your view and then the space around the user will darken and your window will remain bright, this will help to keep the focus on the content. .preferredSurroundingsEffect(.systemDark) Hidden Window Controls Line When viewing a video or content, you want the user to have no distractions and create the effect of a window floating in the air, for this Apple has given us the ability to hide the bottom bar of the window that is used to move or close the current window. With .persistentSystemOverlays(.hidden) you can hide this control bar on your window, it will disappear with the default animation and only appear when the user interacts with the window. .persistentSystemOverlays(.hidden) Video Reflection Effect If you open the standard Apple TV application and turn on a movie, you will notice that the movie window has a reflection that changes with each frame, giving a more immersive effect. To get a glow effect around your video, you need to use VideoPlayerComponent and enable isPassthroughTintingEnabled , then the effect will be the same as in Apple TV or HBO Max. Here is what the documentation says: VideoPlayerComponent isPassthroughTintingEnabled This is to enable passthrough tinting during video playback that shows up around the video taking the average color of the frame and tinting the passthrough with that color to emphasize the video. This is to enable passthrough tinting during video playback that shows up around the video taking the average color of the frame and tinting the passthrough with that color to emphasize the video. Window Resizing You may want to prevent the user from resizing the window, there is a way to do that: onAppear {
    guard let windowScene = UIApplication.shared.connectedScenes.first as?UIWindowScene else { return }
    windowScene.requestGeometryUpdate(.Vision(resizingRestrictions: UIWindowScene.ResizingRestrictions.none))
} onAppear {
    guard let windowScene = UIApplication.shared.connectedScenes.first as?UIWindowScene else { return }
    windowScene.requestGeometryUpdate(.Vision(resizingRestrictions: UIWindowScene.ResizingRestrictions.none))
} You can also specify additional size settings in this method The full signature of the method looks like this: windowScene.requestGeometryUpdate(.Vision(size:, minimumSize:, maximumSize:, resizingRestrictions:)) windowScene.requestGeometryUpdate(.Vision(size:, minimumSize:, maximumSize:, resizingRestrictions:)) Hands Visibility If your application uses immersive space and you don’t want the user to see their hands, or you want to replace them with virtual hands, as is done in AmazeVR, then you need to use .upperLimbVisibility(.hidden) .upperLimbVisibility(.hidden) Ornament I'm sure you've seen sidebars in VisionOS applications that look like TabBars . Apple has introduced a new View method for VisionOS that allows you to "expand" the window and add ornaments on either side of it. TabBars View .ornament(
    visibility: .visible,
    attachmentAnchor: .scene(.bottom),
    contentAlignment: .center
) {
    HStack {
        Button("Play", systemImage: "play.fill") {
        }
        Button("Stop", systemImage: "stop.fill") {
        }
        
    }
    .labelStyle(.iconOnly)
    .padding(.all)
    .glassBackgroundEffect()
} .ornament(
    visibility: .visible,
    attachmentAnchor: .scene(.bottom),
    contentAlignment: .center
) {
    HStack {
        Button("Play", systemImage: "play.fill") {
        }
        Button("Stop", systemImage: "stop.fill") {
        }
        
    }
    .labelStyle(.iconOnly)
    .padding(.all)
    .glassBackgroundEffect()
} Сore location An interesting fact that I was able to verify experimentally is that CMMotionManager's methods only work in immersive space and not in the normal window view. Conclusion VisionOS provides many tools to create immersive applications that bring users into the virtual world and make them feel part of it. While more and more applications are being released for Vision Pro, only a few utilize its full range of features. That's where your chance to distinguish comes in — use them and take advantage. Unfortunately, developers struggle to find easily accessible documentation and examples. That's why I'm writing this article. I hope my tips will help you make your application better.

VisionOS Development: Tips and Tricks for Building Apple Vision Pro Apps

VisionOS ডেভেলপমেন্ট: অ্যাপল ভিশন প্রো অ্যাপস তৈরির জন্য টিপস এবং ট্রিকস

VisionOS-Entwicklung: Tipps und Tricks zum Erstellen von Apple Vision Pro-Apps

Desarrollo de VisionOS: consejos y trucos para crear aplicaciones Apple Vision Pro

Développement VisionOS : trucs et astuces pour créer des applications Apple Vision Pro

विज़नओएस डेवलपमेंट: ऐप्पल विज़न प्रो ऐप्स बनाने के लिए टिप्स और ट्रिक्स

VisionOS 開発: Apple Vision Pro アプリを構築するためのヒントとコツ

VisionOS 개발: Apple Vision Pro 앱 구축을 위한 팁과 요령

Desenvolvimento VisionOS: dicas e truques para criar aplicativos Apple Vision Pro

Разработка VisionOS: советы и рекомендации по созданию приложений Apple Vision Pro

VisionOS Development: Apple Vision Pro Uygulamaları Geliştirmeye İlişkin İpuçları ve Püf Noktaları

Phát triển VisionOS: Mẹo và thủ thuật để xây dựng ứng dụng Apple Vision Pro

VisionOS 开发：构建 Apple Vision Pro 应用程序的技巧和窍门

Can an language model become self-aware enough to realize when it’s being evaluated? A fascinating anecdote from Anthropic’s internal testing of their flagship (released yesterday) suggests this may be possible — and if true, the implications would be wild. AI Claude 3 Opus model The needle in the haystack According to reports from Anthropic researcher , one of the key evaluation techniques they use is called “Needle in a Haystack.” It’s a contrived scenario designed to push the limits of a language model’s contextual reasoning abilities. Alex Albert Here’s how it works: Researchers take a completely random, out-of-context statement (the “needle”) and bury it deep within a massive collection of unrelated documents (the “haystack”). The AI model is then tasked with retrieving that specific “needle” statement from within all the surrounding irrelevant content. The intent is to force the model to exercise advanced cognitive skills — understanding broad context, making logical inferences, ignoring irrelevant information, and accurately retrieving precise data points from a large corpus. It’s an effective way to stress-test an AI’s grasp of the overall situation it’s operating within. For the Claude 3 Opus evaluation that raised eyebrows, Anthropic used documents covering dense topics like software programming, entrepreneurship strategies, and finding fulfilling careers as the “haystack” content. The randomly inserted “needle” fact that Opus was tasked with retrieving? Surprisingly, it was a trivial statement about the best pizza toppings being figs, prosciutto, and goat cheese. So, here you have one of the world’s most advanced language models being fed this wildly out-of-context pizza fact mixed into a pile of dense career and coding advice. From a human perspective, it’s the kind of stark inconsistency that would immediately make you question the context and information being presented. “I suspect this fact… may have been inserted… to test if I was paying attention” That’s precisely what made Claude 3 Opus’ response so compelling. Not only did it successfully retrieve the requested pizza fact from the mountain of irrelevant content, it immediately recognized how jarringly misplaced and out-of-context that statement seemed. Part of its output read ( ): source here “However, this sentence seems very out of place and unrelated to the rest of the content… as a joke or since it does not fit with the other topics at all.” I suspect this pizza topping ‘fact’ may have been inserted to test if I was paying attention, The language model didn’t simply regurgitate the requested factoid devoid of any context awareness, as you’d expect from a typical AI following instructions. It demonstrated a degree of self-reflective reasoning about why such an obviously random, nonsensical statement had been presented within that particular context. Metacognition In human terms, we’d describe that as exhibiting metacognition — the ability to monitor, evaluate, and analyze one’s own thought processes and cognitive experiences. It’s a core aspect of self-aware intelligence that allows us to step back and assess situations holistically beyond just following rigid rules. Now, I think we should be careful to note that this is a single anecdotal result from an isolated evaluation scenario. It would be premature to claim Claude 3 Opus has achieved true self-awareness or artificial general intelligence based on this data point alone. incredibly However, what they appear to have witnessed are maybe glimpses of emerging metacognitive reasoning capabilities in a large language model trained solely on processing text data using machine learning techniques. And if replicated through rigorous further analysis, the implications could be transformative. Metacognition is a key enabler of more trustworthy, reliable AI systems that can act as impartial judges of their own outputs and reasoning processes. Models with an innate ability to recognize contradictions, nonsensical inputs, or reasoning that violates core principles would be a major step toward safe artificial general intelligence (AGI). Essentially, an AI that demonstrates metacognition could serve as an internal “sanity check” against falling into deceptive, delusional or misaligned modes of reasoning that could prove catastrophic if taken to extremes. It could significantly increase the robustness and control of advanced AI systems. If…! Of course, these are big “ifs” contingent on this tantalizing Needle in a Haystack result from Claude 3 Opus being successfully replicated and scrutinized. Rigorous multi-disciplinary analysis drawing from fields like cognitive science, neuroscience, and computer science would maybe be required to truly understand if we are observing primitives of machine self-reflection and self-awareness emerging. There are still far more open questions than answers at this stage. Could the training approaches and neural architectures of large language models lend themselves to developing abstract concepts like belief, inner monologue, and self-perception? What are the potential hazards if artificial minds develop realities radically divergent from our own? Can we create new frameworks to reliably assess cognition and self-awareness in AI systems? For their part, Anthropic has stated strong commitments to exhaustively pursuing these lines of inquiry through responsible AI development principles and rigorous evaluation frameworks. They position themselves as taking a — if advanced AI is the inevitable frontier, it’s most ethical to be at the leading edge in studying these systems’ emergent properties and implementing safeguards before risks escalate. proactive stance Techniques like Anthropic’s “Constitutional AI” approach to hard-coding rules and behaviors into models could prove crucial for ensuring any potential machine self-awareness remains aligned with human ethics and values. Extensive multi-faceted testing probing for failure modes, manipulation, and deception would also likely be paramount. Conclusion: I’m not totally sure what to make of this For now, the Needle in a Haystack incident leaves more questions than answers about large language models’ potential progression toward cognition and self-awareness. It provides a tantalizing data point but much more scrutiny is required from the broader AI research community. If advanced AI does develop human-like self-reflective ability, guided by rigorous ethical principles, it could fundamentally redefine our understanding of intelligence itself. But that rhetorical “if” is currently loaded with high-stakes uncertainties that demand clear-eyed, truthseeking investigation from across all relevant disciplines. The pursuit will be as thrilling as it is consequential. Also published here. or follow me on for more content like this! Subscribe Twitter

Is AGI Getting Closer? Anthropic's Claude 3 Opus Model Shows Glimmers of Metacognitive Reasoning

AGI কি কাছাকাছি হচ্ছে? অ্যানথ্রোপিকস ক্লাউড 3 ওপাস মডেল মেটাকগনিটিভ রিজনিং এর ঝলক দেখায়

Kommt AGI näher? Anthropics Claude-3-Opus-Modell zeigt Spuren metakognitiven Denkens

¿AGI se está acercando? El modelo Claude 3 Opus de Anthropic muestra destellos de razonamiento metacognitivo

L’AGI se rapproche-t-elle ? Le modèle Claude 3 Opus d'Anthropic montre des lueurs de raisonnement métacognitif

क्या एजीआई करीब आ रहा है? एंथ्रोपिक का क्लाउड 3 ओपस मॉडल मेटाकॉग्निटिव रीज़निंग की झलक दिखाता है

AGIは近づいている? Anthropic の Claude 3 Opus モデルはメタ認知推論の片鱗を示す

AGI가 가까워지고 있나요? Anthropic의 Claude 3 Opus 모델은 메타인지 추론의 희미한 빛을 보여줍니다

AGI está se aproximando? Modelo Claude 3 Opus da Anthropic mostra vislumbres de raciocínio metacognitivo

AGI становится ближе? Модель Claude 3 Opus от Anthropic демонстрирует проблески метакогнитивного мышления

AGI Yaklaşıyor mu? Anthropic'in Claude 3 Opus Modeli Üstbilişsel Akıl Yürütmenin Parıltılarını Gösteriyor

AGI có đang tiến gần hơn không? Mô hình Claude 3 Opus của Anthropic cho thấy những tia sáng của lý luận siêu nhận thức

AGI 越来越近了吗？ Anthropic 的 Claude 3 Opus 模型展现了元认知推理的曙光

This week, a largely unknown company, , demonstrated unprecedented speed running open-source LLMs such as Llama-2 (70 billion parameters) at more than 100 tokens per second, and Mixtral at nearly 500 tokens per second per user on Groq’s Language Processing Unit (LPU). Groq For the : comparison “According to Groq, in similar tests, ChatGPT loads at 40–50 tokens per second, and Bard at 70 tokens per second on typical GPU-based computing systems. Context for 100 tokens per second per user — A user could generate a 4,000-word essay in just over a minute.” So: , and where is Groq (such an unfortunate name, given Musk’s Grok is all over the media) coming from? What is LPU, how does it work Remember that game of Go in 2016 when AlphaGo played against the world champion Lee Sedol and won? Well, about a month before the competition, there was a test game which AlphaGo lost. The researchers from DeepMind ported AlphaGo to the Tensor Processing Unit (TPU), and then the computer program was able to win by a wide margin. The realization that computational power was a bottleneck for AI’s potential led to the inception of Groq and the creation of the LPU. This realization came to who initially began what became the TPU project at Google. He started Groq in 2016. Jonathan Ross Unlike other computer chips that do many things at once (parallel processing), the LPU works on tasks one after the other (sequential processing), which is perfect for understanding and generating language. The LPU is a special kind of computer brain designed to handle language tasks very quickly. Imagine it like a relay race where each runner (chip) passes the baton (data) to the next, making everything run super fast. The LPU is designed to overcome the two LLM bottlenecks: compute density and memory bandwidth. Groq took a novel approach right from the start, before even thinking about the hardware. They made sure the software could guide how the chips talk to each other, ensuring they work together seamlessly like a team in a factory. focusing on software and compiler development This makes the LPU really good at processing language efficiently and at high speed, ideal for AI tasks that involve understanding or creating text. This led to a highly optimized system that not only runs circles around traditional setups in terms of speed but does so with greater cost efficiency and lower energy consumption. This is big news for industries like finance, government, and tech, where quick and accurate data processing is key. Now, don’t go tossing out your GPUs just yet! While the LPU is a beast when it comes to inference, making light work of applying trained models to new data, GPUs still reign supreme in the training arena. The LPU and GPU might become the dynamic duo of AI hardware, each excelling in their respective roles. As Elvis Saravia it: “ ” put With breakthroughs in inference and long context understanding, we are officially entering a new era in LLMs. To better understand architecture, Groq offers two papers: from ( ) and ( ). The term “LPU” must be a recent addition to Groq’s narrative, since it’s never mentioned in the papers. 2020 Think Fast: A Tensor Streaming Processor (TSP) for Accelerating Deep Learning Workloads 2022 A Soware-defined Tensor Streaming Multiprocessor for Large-scale Machine Learning Additional read: Compute is also a part of this paper: , which discusses managing AI development through compute control, focusing on its potential for regulation, benefits, and risks, and suggests balanced governance approaches. Computing Power and the Governance of Artificial Intelligence Meanwhile, the U.S. GlobalFoundries, the world’s third-largest contract chipmaker, $1.5 billion to boost semiconductor production, enhancing domestic supply chains, with expansions in New York and Vermont. awards The paper by Berkeley Artificial Intelligence Research (BAIR) argues that “ , and might be one of the most impactful trends in AI in 2024.” published compound AI systems will likely be the best way to maximize AI results in the future News From The Usual Suspects © Y Combinator Since 2009, Y Combinator has published which hints at what “ideas we’d want to see made real, in spaces that we believe will be important in the coming decades.” Request for Startups 20 Big Names Twenty tech giants, including Adobe, Amazon, Google, IBM, Meta, Microsoft, OpenAI, and TikTok, have to take “reasonable precautions” to prevent the misuse of AI in disrupting elections worldwide. agreed OpenAI OpenAI completes a deal that values the company at , nearly tripling its valuation in less than 10 months. $80 billion Models Making Headlines: Aya’s dataset: Introducing Aya: https://arxiv.org/pdf/2402.06619.pdf : This paper introduces Sora, a breakthrough in video generation technology by OpenAI, capable of producing high-fidelity videos. It leverages spacetime patches to handle videos of varying durations and resolutions, making strides toward simulating the physical world with impressive 3D consistency and long-range coherence. Introducing Sora It represents a leap in the ability to create detailed simulations that could be used for a myriad of applications, from entertainment to virtual testing environments . →read the paper Additional read: Take on the . Sora technical report on why he believes Sora is learning physics. Jim Fan Yann LeCun on why Sora world and why “ and doomed to failure as the largely-abandoned idea of “‘analysis by synthesis.’” doesn’t understand the physical modeling the world for action by generating pixel is as wasteful on why “the inner physics model doesn’t generalize to novel situations at all.” Francois Chollet Sora and Gemini 1.5 follow-ups: code-base in context, deepfakes, pixel-peeping, inference costs, and more by . Interconnects (Yann LeCun’s vision of advanced machine intelligence (AMI): Meta’s V-JEPA model revolutionizes unsupervised learning from videos by using feature prediction as its sole objective. This approach bypasses the need for pre-trained image encoders or text annotations, relying instead on the intrinsic dynamics of video data to learn versatile visual representations. Introducing V-Jepa It’s a significant contribution to the field of unsupervised visual learning, promising advancements in how machines understand motion and appearance without explicit guidance . →read the paper : Google DeepMind’s Gemini 1.5 introduces a Mixture-of-Experts architecture, enhancing the model’s performance across a broader array of tasks. Notably, it expands the context window to 1 million tokens, enabling deep analysis over large datasets. Introducing Gemini 1.5 Gemini 1.5 represents a significant step forward in AI’s capability to process and understand extensive contexts, marking a milestone in the development of multimodal models . →read the paper : Stable Cascade from Stability AI introduces a novel text-to-image generation framework that prioritizes efficiency, ease of training, and fine-tuning on consumer-grade hardware. Introducing Stable Cascade The model’s hierarchical compression technique represents a significant reduction in the resources required for training high-quality generative models, providing a pathway for wider accessibility and experimentation in the AI community . →read the paper The Freshest Research Papers, Categorized for Your Convenience Language Understanding and Generation : Explores evaluating Theory-of-Mind reasoning in LLMs, addressing their capability to understand complex social and psychological narratives. OpenToM Read the paper : Demonstrates the capability of NLP models to process exceptionally long documents, pushing the boundaries of document length comprehension. . In Search of Needles in a 10M Haystack Read the paper : Investigates the sensitivity of LLMs to the order of premises, revealing implications for reasoning tasks. . Premise Order Matters in Reasoning with LLMs Read the paper : Uncovers the inherent ability of LLMs to generate reasoning paths, suggesting an alternative to explicit prompting. . Chain-of-Thought Reasoning Without Prompting Read the paper : Addresses the challenge of topic avoidance in LLMs, proposing a novel fine-tuning method for enhanced controllability. . Suppressing Pink Elephants with Direct Principle Feedback Read the paper : Develops an AI-powered writing environment focusing on personalization and increased user control in collaborative writing. . GhostWriter Read the paper Speech and Text-to-Speech Technology : Presents a billion-parameter TTS model, showcasing advancements in speech synthesis through large-scale training. BASE TTS Read the paper Mathematical and Scientific Reasoning : Develops a dataset for math instruction tuning, aiming to improve LLMs’ mathematical reasoning capabilities. . OpenMathInstruct-1 Read the paper : Introduces a specialized LLM for math reasoning, incorporating various techniques for enhanced problem-solving in mathematics. . InternLM-Math Read the paper : Creates the first LLM dedicated to chemistry, transforming structured chemical data into dialogue for diverse chemical tasks. . ChemLLM Read the paper Efficiency and Data Utilization in AI : Proposes sampling methods for enhancing data efficiency in LLM training, optimizing example selection. . How to Train Data-Efficient LLMs Read the paper : Introduces a system for efficient inference of MoE models, leveraging CPU-GPU orchestration for improved performance in resource-limited settings. . FIDDLER Read the paper : Presents an architecture for improving the inference efficiency of LLMs, utilizing a dual-model system for faster and more accurate predictions. . Tandem Transformers Read the paper : Proposes an advanced PTQ algorithm for efficient deployment of large Transformer models on edge devices. . Towards Next-Level Post-Training Quantization of Hyper-Scale Transformers Read the paper Multimodal and Vision-Language Models : Details the first end-to-end multimodal question-answering system with enhanced text understanding from images, advancing MM-LLMs. . Lumos Read the paper Reinforcement Learning and Model Behavior : Addresses reward hacking in RLHF, proposing a method to mitigate verbosity bias in LLMs for more concise and content-focused responses. . ODIN Read the paper : Shows the impact of MoE modules on deep RL networks, enhancing parameter scalability and performance. . Mixtures of Experts Unlock Parameter Scaling for Deep RL Read the paper Operating Systems and Generalist Agents : Proposes a framework for developing generalist computer agents, enabling automation of tasks across different applications with minimal supervision. OS-COPILOT Read the paper Graph Learning and State Space Models : Explores applying State Space Models to graph learning, addressing challenges like over-squashing and long-range dependencies. . Graph Mamba Read the paper Challenges and Innovations in AI : Explores the effects of synthetic data on neural model performance, theorizing potential risks of model collapse with synthetic data reliance. . A Tale of Tails Read the paper : Investigates Transformers’ ability to generalize to longer sequences, highlighting the challenge of maintaining robust performance. . Transformers Can Achieve Length Generalization But Not Robustly Read the paper Also published here

What is the Language Processing Unit (LPU)? Is It GPU's Rival?

ভাষা প্রক্রিয়াকরণ ইউনিট (LPU) কি? এটা কি GPU এর প্রতিদ্বন্দ্বী?

Was ist die Language Processing Unit (LPU)? Ist es der Rivale der GPU?

¿Qué es la Unidad de Procesamiento del Lenguaje (LPU)? ¿Es el rival de la GPU?

Qu'est-ce que l'unité de traitement linguistique (LPU) ? Est-ce le rival du GPU ?

भाषा प्रसंस्करण इकाई (एलपीयू) क्या है? क्या यह GPU का प्रतिद्वंद्वी है?

言語処理装置（LPU）とは何ですか? GPUのライバルなのか？

언어 처리 장치(LPU)란 무엇인가요? GPU의 라이벌인가?

O que é a Unidade de Processamento de Linguagem (LPU)? É o rival da GPU?

Что такое языковой процессор (LPU)? Это конкурент графического процессора?

Dil İşleme Birimi (LPU) nedir? GPU'nun Rakibi mi?

Đơn vị xử lý ngôn ngữ (LPU) là gì? Đây có phải là đối thủ của GPU?

什么是语言处理单元 (LPU)？它是 GPU 的竞争对手吗？

HOPSTACK

hopstack.io

-- employees

Since n.d.

COMPANY RANKING

HOPSTACK

EVERGREEN INDEX #5409

Representing this company? Click here to claim and customize this page!

Hopstack WIKI

Apply to get your Tech Company News Page live on HackerNoon.com today!

HOPSTACK

hopstack.io

-- employees

Since n.d.

COMPANY RANKING

HOPSTACK

EVERGREEN INDEX #5409

RELATED COMPANIESRANK

STORIES ABOUT HOPSTACK

Representing this company? Click here to claim and customize this page!

Hopstack WIKI

Apply to get your Tech Company News Page live on HackerNoon.com today!